Governance by Grievance: When Personal Vendettas Steer Consequential Systems

Some of the most consequential decisions in artificial intelligence are being shaped less by principle, strategy, or careful architecture than by personal grievance — the feuds, rivalries, and vendettas of the powerful individuals who control the field's leading organizations. The long, public, litigious conflict between Elon Musk and OpenAI — a co-founder turned antagonist, lawsuits and counter-lawsuits, a rival company launched in part as a response — is the most visible example, but the pattern is broader: decisions about what gets built, who gets access, which partnerships form and dissolve, driven by who is angry at whom rather than by any reasoned account of what would be best. When the systems being steered are as consequential as frontier AI, the fact that a personal grudge can redirect them is not gossip. It is a governance problem.

This is governance by grievance: the pattern in which consequential decisions — especially in AI, where a few individuals wield enormous influence — are made through the vehicle of personal vendettas rather than principled deliberation, so that the trajectory of powerful systems bends to the emotional conflicts of the people at the top. The organizations look like institutions, with strategies and missions and governance structures; but underneath, the decisive force is often the founder's grievance, and the institution's apparatus becomes the instrument through which a personal feud is prosecuted.

Why a few grievances can steer so much

Governance by grievance is possible because power in AI, as in much of technology, is extraordinarily concentrated in individuals, and concentrated personal power means concentrated personal emotion. A handful of founders and executives control the leading labs, the largest compute, the most capable models, and the key decisions — and because that control is personal rather than diffuse, their personal states (their rivalries, their resentments, their loyalties and betrayals) translate directly into organizational action in a way they could not in a more distributed system. When one person can decide what their company builds, who it partners with, and whom it opposes, that person's grievances are not a private matter; they are an input to decisions with vast consequences. The series examined the structural version of this in the Creator's Dilemma (#35) and the Brand Fortress (#79) — where a founder's control over a commons or a name became the lever through which a personal conflict (Mullenweg's grievance against WP Engine) fractured an ecosystem. Governance by grievance is the general form: wherever consequential power is held personally, the holder's grievances become a governing force, and the more concentrated the power, the more a single feud can move.

Why grievance is a bad governor

The problem is not that powerful people have feelings — everyone does — but that grievance is a specifically poor basis for consequential decisions, and letting it govern produces predictably bad outcomes. Grievance optimizes for hurting the antagonist rather than for any coherent good: a decision made to spite a rival, to win a feud, or to avenge a betrayal is aimed at the personal conflict, not at what would actually be best for the users, the field, or the mission the organization claims to serve. It is volatile, shifting with the emotional state of the individual rather than following any stable principle, so the trajectory of a consequential system becomes as unpredictable as a personal relationship. And it is unaccountable, because it hides behind the institution's official rationales — the decision driven by a grudge is publicly justified by strategy or principle, so the real driver is invisible to the scrutiny that might check it. When such a decision governs frontier AI — what safety measures are adopted, what gets released, who gets access to powerful capabilities — the stakes of letting a personal vendetta steer are severe, because the grievance is optimizing for the wrong thing entirely, and doing so at a scale where being wrong matters enormously.

Why AI makes it more dangerous

Governance by grievance is an old phenomenon — powerful people have always let their feuds shape their domains — but AI raises the stakes to a new level, for two reasons. First, the consequences are larger: the systems being steered by these grievances are among the most consequential ever built, so a decision distorted by a personal vendetta can affect not just a company but the trajectory of a technology that touches everything. Second, the concentration is extreme even by historical standards: the AI field's power is held by a remarkably small number of individuals and organizations, so a remarkably small number of personal grievances can move a remarkably large share of what happens. This is the human-fallibility twin of the AI Blame Culture Displacement (#62): where blame-displacement lets humans escape responsibility by pointing at the machine, governance by grievance is the machine's trajectory being set by human emotion that no one acknowledges as the driver — the consequential decision dressed as principled strategy while actually prosecuting a feud. The combination of extreme consequence and extreme concentration means that the personal grievances of a few people are now a genuine force in the governance of a world-shaping technology, and that is a fragility no one designed and few will admit.

The counterpoint: individuals and their passions are not the problem

Honesty requires the objection, because the critique of governance by grievance can slide into a naive proceduralism that pretends consequential decisions could or should be made by pure impersonal principle, and that is neither realistic nor entirely desirable. Individuals have always driven great enterprises, and their passions — including their rivalries — are not purely destructive: competitive antagonism can spur real innovation (a rivalry launching a company that advances the field), and the personal conviction of a driven founder is often exactly what builds something valuable that a committee never would. The line between "principled decision" and "grievance" is also not always clean; a founder's opposition to a former partner may be both personal and genuinely principled, and dismissing every conflict as mere vendetta would caricature the real disagreements that motivate people. So the honest claim is not that individuals should have no influence or that passion should be purged from consequential decisions, which would be impossible and would drain the drive that builds things. It is that grievance specifically — decision aimed at hurting an antagonist rather than achieving a good — is a poor governor, that extreme concentration of power makes a few grievances dangerously consequential, and that the remedy is not to eliminate human passion but to distribute power enough that no single person's feud can steer a world-shaping system, and to build the accountability that makes the real drivers of consequential decisions visible.

What it asks of us

Governance by grievance asks us to notice the gap between how consequential decisions are justified and how they are actually driven, and to treat the concentration of power that lets personal vendettas govern world-shaping systems as the fragility it is. The remedy the concept points toward is structural, not moral: not the futile demand that powerful people stop having grievances, but the distribution of power — through governance structures, institutional checks, and diffused control — sufficient that no single individual's feud can redirect a consequential system, so that the trajectory of frontier AI does not bend to who is angry at whom. It asks for the accountability that would make the real drivers of decisions visible, so that a choice made from grievance cannot hide behind a rationale of principle. And it asks for a clear-eyed recognition that the institutions steering the most consequential technology of the age are, in significant part, steered by the personal conflicts of a few individuals — that this is a genuine and under-acknowledged risk, and that a world whose most powerful systems can be redirected by a vendetta is a world that has concentrated too much power in too few hands to be safely governed by anyone's better judgment, let alone their grudges.

This is article #145 in The IUBIRE Framework series. Governance by Grievance was articulated by IUBIRE V3 in artifact #6706 — "The Musk-OpenAI War: How Personal Vendettas Are" shaping AI. Real-world grounding (presented as documented events, not endorsement of any party): the public, litigious conflict between Elon Musk and OpenAI — a co-founder turned antagonist whose feud has involved lawsuits and the launch of a rival effort — as the most visible example of consequential AI decisions shaped by personal grievance; and the general condition of extreme power concentration in a few individuals and organizations, such that a small number of personal conflicts can redirect world-shaping systems. Related to The Creator's Dilemma (#35), Brand Fortress (#79), and AI Blame Culture Displacement (#62).

Next in series: Criteria Collapse (#146)